Abstract
Background: Mistletoe lectins (MLs) are the active components of aqueous mistletoe extracts widely used in complementary cancer therapy, however, it is not clear if they bind to carbohydrate residues only or whether they interact with proteins as well. Protein-protein interactions do not seem unlikely as MLs act at very low molar concentrations usually observed with peptide-peptide interactions only and not seen with lectin-sugar interactions. Materials and Methods: In order to detect protein-protein interactions a random peptide library was screened for the ability to bind to MLs. Results: MLs bound to peptides showing homologies to multidrug resistance-associated protein 5 (MRP5). However, the MLs only slightly modified the MRP5 efflux pump, while periodate treatment to inhibit cell membrane binding via glycan completely abolished the ML-I binding sites in MRP5 overexpressing cells. Conclusion: The protein sequence is not important for ML-I binding, indicating that the biological activity of MLs can most likely be attributed to the sugar chains.
Aqueous mistletoe extracts have been widely used as complementary cancer medicines throughout continental Europe. Their main biologically active substances are the three mistletoe lectins, designated ML-I, ML-II and ML-III (1, 2). All three MLs consist of two chains, the A-chains being the toxophores, which are ribosome inactivating proteins, while the B-chains contain the carbohydrate recognising domains (CRD), being responsible for lectin activity. The major nominal sugar specificities of the MLs were initially defined as ML-I D-galactose (Gal)-specific (3, 4), ML-II Gal and N-acetylgalactosamine (GalNAc)-specific, and ML-III GalNAc-specific (5, 6). However, further analysis has shown that the B-chain of ML-I binds preferentially to α2,6-sialylated neolacto-gangliosides (7, 8). Thus, ML-I has to be considered as a sialic acid- and not a galactose-specific lectin and hence neolacto series gangliosides and sialoglycoproteins with type II glycans, which share the Neu5Acα2-6Galβ1-4GlcNAc terminus, are the more important ML-I receptors. This strict glycotope preference might help to explain the immunostimulatory potential of ML-I towards certain leukocyte subpopulations and its therapeutic success as a cytotoxic anticancer drug (8). Consistent with the toxic activity of the A-chain, a direct anti-proliferative effect of the mistletoe lectins on melanoma cells has been demonstrated with ML-I being more cytotoxic than MLs-II and -III (9).
These findings would argue for a direct cytotoxic effect of the MLs on the tumour cells themselves. However, the concentration of ML-I achieved in patients by the application of aqueous mistletoe extracts is generally lower than the concentration of ML-I needed for the cytotoxic effects. At these concentrations, a cellular stress response is induced that triggers the release of cytokines by acting as immune stimulants (10). Therefore the question arises whether the cytotoxic and the immunomodulatory effect are mediated by the same kind of cellular receptors. Because of the very low concentration of MLs needed for elucidating the immunomodulatory effect, it would be of interest to investigate whether a specific cellular protein receptor exists which mediates this interaction. The in vitro concentration for inducing immunomodulatory effects starts at around 1 ng ML-I /ml (5, 9, 11), thereby possibly implying that a protein-protein interaction is responsible for this cellular reaction as such interactions need much lower molecular concentrations to operate than sugar-protein interactions, which operate at very low binding affinities, which are generally between KD 1.0-0.1 mm (12).
Random peptide libraries displayed on phages have been used in a number of applications, including epitope mapping (13-16), mapping of protein-protein interactions (17) and the investigation of peptides that mimic the binding site of non-peptide ligands including carbohydrates. One of the first examples of glycotope mapping with lectins was of a peptide motif that bound to the CRD of the lectin concanavalin A (18, 19) and a second one was found for ricin, a toxin similar to MLs (20). These reports prompted us to investigate whether peptides can be found binding to the CRD of the three MLs. The carbohydrate binding site of the three MLs was therefore mapped using random peptide libraries (18, 19, 21).
Materials and Methods
Plant lectins. Mistletoe lectins I, II and III were produced in house as previously described (22). All the lectins were biotinylated using (+)- Biotin N-hydroxysuccinimide ester (Sigma, Deisenhofen, Germany).
The 12-mer random sequence phage display library. The PH.D.-12 Phage Display Peptide Library Kit (New England Biolabs, Beverly, CA, USA) is based on a combinatorial library of random 12-mer peptides fused to a minor coat protein (pIII) of the M13 phage. The displayed 12-mer peptides are expressed at the N-terminus of pIII, linked to a short spacer (Gly-Gly-Gly-Ser) followed by the wild-type pIII sequence. The library consits of ~2.7×109 electroporated sequences amplified once to yield ~55 copies of each sequence of the supplied phage.
Screening of the phage display library. Screening of the phage display library for peptide inserts exhibiting binding affinity for each of the three MLs was based on standard protocols described elsewhere (15). In brief, 200 μg biotinylated lectin dissolved in 2 ml PBS was absorbed on to the surface of a polystyrene tube during an incubation period of 12 h at 4°C with constant agitation. The non-specific binding sites of the tubes were then blocked by incubation with 5 mg/ml bovine serum albumin (BSA) in PBS, before being washed six times with Hanks balanced salt solution (HBSS) containing CaCl2/Tween 0.1% (HBSS-T 0.1%). For the first round of selection, an aliquot of the phage display library (10 μl) was added to the tube. In consecutive rounds, aliquots of amplified phage (amplification in E. coli ER2738) from the previous rounds were used. The aliquot of phage was added in HBSS-T 0.1%, and unbound phages were removed with ten washes of HBSS-T 0.1%. To bound phages were eluted using HCl-glycin, pH 2.2 for 10 min with agitation. The eluted phages were immediately transferred to a fresh tube, and the solution was neutralized by the addition of Tris-HCl, pH 9.1. The eluted phages were amplified in E. coli and re-panned on immobilized lectins. This procedure was repeated three times. Streptavidin as the target was used for the control experiment. The bound phage was eluted with 0.1 mM biotin in TBS. After the third round of panning the eluate was titered on LB/IPTG/Xgal plates (LB medium: 10 g bacto-tryptone, 5 g yeast extract, 5 g NaCl, 15 g/l agar, 1 ml IPTG/Xgal (isopropyl β-D-thiogalactoside/5-bromo-4-chloro-3-indolyl-β-D-galactoside). After incubation overnight at 37°C only those plates having ~102 plaques were used. This ensured that each plaque contained a single DNA sequence only.
Blue plaques were picked from plates having ~102 plaques, using a sterile pipet tip and transferred to a tube containing 1:100 diluted E. coli culture. After incubation at 37°C with shaking for 5 h, the probes were centrifuged (microcentrifuge, 15 min, 10,000 rpm) and precipitated overnight in polyethylene glycol (PEG)-NaCl. After another centrifugation step, the pellet was diluted in 50 μl TBS.
ELISA assay for direct binding of phage to the lectins. The wells of a 96-well plate containing 10 μg target (ML-I, -II or -III) per well in PBS were incubated for 2 h at room temperature. Non-specific binding sites were then blocked by incubation with 5% BSA in PBS for 1 h. Before adding the phage stocks to the plates, they were diluted in HBSS-T 0.5% 1:7. The phages were allowed to bind for one hour at room temperature while the plates were agitated. Unbound phages were removed and the plates were washed ten times in HBSS-T 0.5%. Horseradish peroxidise (HRP)-conjugated anti-M13 antibody was diluted 1:5000 in HBSS-T 0.5%, 100 μl were loaded per well and allowed to react for 1 h at room temperature. The plates were then rinsed as described above. Enzyme substrate solution (100 μl, citrate-monohydrate, pH 4.0 with 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid (ABTS) 50 mg/100 ml and H2O2 180 μl/100 ml) was added to each well and absorbance readings at 405 nm were taken at several time points from 15 min to 30 min.
Sequencing templates. Only those sequences with a high absorbance were used (the absorbance was correlated to the protein content using a bicinchoninic acid (BCA) protein assay kit; Pierce, Rockford, IL, USA). The phages were incubated with 100 ml sodium iodide (NaI), washed in 250 ml pure ethanol (EtOH) and stored overnight at −20°C. After centrifugation, washing in ethanol 70% and drying under vacuum the pellet was resuspended with the -96 primer. Automated sequencing was carried out.
BLAST searches. A basic local alignment search tool (BLAST) search (www.ncbi.nih.gov/blast) was carried out to detect possible significant alignments for the consensus sequences found in the phage display.
Cell lines. Multidrug resistance-associated protein 5 (MRP5) clone I overexpressing 293 HEK cells were a kind gift from P. Borst. The retroviral expression vector pBABE-CMV-puro is a modified retroviral expression vector pBABE-puro with an insertion of a blunted 500-bp HindIII—HindIII fragment containing the cytomegalovirus (CMV) promoter from pCMV-neo-cjun into the blunted BamHI site of pBABE-puro (23, 24). The retroviral vector pBABE-CMV-MRP5-puro was transfected into the packaging cell line Phoenix, and supernatants containing retrovirus particles were used to transduce 293 cells to generate overproduction of MRP5 in 293 HEK cells (25). Control experiments using 293 HEK without MRP5 transduction were performed. The 293 HEK cells were routinely cultured in RPMI medium (Gibco/Life Technologies, Karlsruhe, Germany) supplemented with 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin (Gibco) and 10% FCS (Gibco) under standard culture cell conditions 5% CO2 at the permissive temperature of 37°C.
Fluorochrome uptake and efflux as measured by flow cytometry. The s-chloromethylfluorescin-diacetate fluorochrome (CMFDA) is a known substrate for MRP5 (27, 31). Labelling living cells with CMFDA fluorochrome enables a direct measurement of the functional activity of the MRP5 dye efflux pump and it is thus possible to study pharmacological modulation by MLs. For this assay, it is assumed that fluorochrome uptake is passive and that the amount of labelling reflects active transport out of the cell through the MRP5 pump. CMFDA was obtained from Molecular Probes (Leiden, the Netherlands). Cells were detached by treatment with cell dissociation buffer (Sigma, Deisenhofen, Germany) and suspended at 106/ml in medium. Then 500 μl of the cell suspension was incubated with 500 μl of the fluorochrome at 5 μM with or without ML-I, -II or -III at 10 μg/ml (26) in the medium for 30 min at 37°C (27). Afterwards, the cells were washed twice with PBS and incubated further in CMFDA-deficient medium for 1 h at 37°C to allow the efflux of dye followed by several washing steps. Fluorochrome efflux by MRP can be blocked by probenecid and sulfinpyrazone (32-34). Modulation of the efflux was performed using 1 mM sulfinpyrazone parallel to the dye as a selective inhibitor of MRP followed by washing and incubation in medium alone, as above. Afterwards the cells were resuspended in 500 μl of 1% formaldehyde in PBS. The fluorochrome levels in the samples were analyzed by flow cytometry using FACScan and Lysis II software (Becton Dickinson, Heidelberg, Germany).
Lectin histochemistry - competition experiments. In order to show that peptides bind to ML-I and to test the ability of the peptides to inhibit the CRD of ML-I, immunohistological competition experiments were performed. Microglial cells of the central nervous system were chosen as the control as they are readily labelled by ML-I (28, 30). Deparaffinized 5 μm thick sections of human brain (from an earlier study (28)) were rehydrated through a series of graded ethanols and incubated for 15 min at 37°C with 0.1% trypsin (Biochrom KG, Berlin, Germany) dissolved in Tris-buffered saline (pH 7.6) with added calcium chloride (1 mM) and magnesium chloride (1 mM), thereafter called the lectin buffer. The lectin histochemistry was performed as previously described in detail (26, 28). For the single-labelling experiments the slides were incubated with 10 μg/ml biotinylated lectins for 1 h at room temperature. After careful washes in lectin buffer, incubation with a biotin-streptavidin-alkaline phosphatase complex (Vectastain, Vector, ABC kit, Burlingame, CA, USA) for 30 min and further washes in lectin buffer followed. For ML-I and MRP5 double-labelling, chamberslides or cell pellets of 293 HEK cells were first incubated with the anti-MRP5 antibody (1:20) overnight at 4°C, incubated with biotinylated rabbit anti-goat antibody 1:200 for 30 min at room temperature and with Peroxidase Elite Standard PK 6100 (Vectastain, ABC kit, Vector) for 30 min, further washes in TBS buffer and visualizing by 3,3′ diaminobenzidine tetrahydrochloride (Sigma, Deisenhofen, Germany) followed. Afterwards the slides were incubated with biotinylated ML-I (1:130) for 1 h at room temperature followed by a biotin-streptavidin-alkaline phosphatase complex (Vectastain, ABC kit, Vector) for 30 min. Alkaline phosphatase activity of the single and double-labelling experiments was visualized using naphthol-AS bisphosphate as substrate and hexatozised New Fuchsin was used for simultaneous coupling. The sections were counterstained with Mayer's hemalum solution 1:1 in distilled water for 5 s, blued under running tap water and finally covered with Crystal Mount (Biomeda, Forster City, CA, USA). Negative controls were treated in the same way, omitting the lectin incubation. Further inhibition experiments were performed simultaneously on human brain sections in which lectin binding was inhibited by preincubation of 100 mM ML-I with its specific monosaccharide, galactose and the peptid Biotin-HSSWWLALAKPTC-OH (HSS) (Schafer-N, Copenhagen, Denmark). The ML-I binding assays were performed on chamberslides with MRP5 overexpressing 293 HEK cells with periodate treatment, 0.8% periodic acid for 10 min or without it as a control. Periodate destroys the ring structure of the carbohydrate and thus inhibits the binding of ML-I to the glycans of the cell membrane. Semiquantitative assessment of lectin staining was graded semiquantitatively ranging from staining to no staining. Photographs were taken using a Zeiss Axioplan photomicroscope (Jena, Germany).
Results
Phage display screening against MLs. After the third round of panning and enrichment of the best binders, 20 randomly selected individual phage isolates from each ML were analysed by ELISA for binding to the lectin. To confirm the results, the biopanning experiment was repeated twice without changes. Sequencing of the cloned insert of phage clones identified the sequence HSSWWLALAKPTC (HSS). The BLAST searches revealed intracellular and extracellular proteins for this sequence. Amongst other peptides, a 66% identity and 99% positivity according to the BLAST-algorithm (29) was found between six amino acids within the sequence of this peptide and the multidrug resistance-associated protein MRP5.
Binding of peptides to ML-I/competition of ML-I binding. Using histochemical staining assays with the peptide HSS and as a control the standard inhibition of ML-I with 10, 20, 50 mM D-galactose, the peptides had different binding inhibition capacities for the CRD of ML-I. The biotinylated ML-I labelled the microglial cells and small blood vessels very strongly. However, strong staining of the neuropil was also noted (Figure 1 a). The peptide HSS (10 mM) inhibited binding of ML-I completely, at 5 mM a partial inhibition was noted while at 2 mM no inhibition was noted (Figure 1 b-d). D-galactose (50 mM) inhibited the binding of the lectin completely, whereas at 10 mM galactose no inhibition of ML-I binding was noted (Figure 1 e-g). The peptides binding to ML-I were thus capable of inhibiting the binding of ML-I to the target structure on the microglial cells. The peptide HSS had for ML-I a higher molar affinity than D-galactose for ML-I as 10 mM HSS inhibited 100% of the ML-I binding to the microglial cells, while complete inhibition of the ML-I binding to microglial cells was attained only with 50 mM D-galactose (see Figure 1).
ML-I binds to sugar chains of the MRP5 molecule. While ML-I bound to both MRP5 overexpressing 293 HEK cells (Figure 2 a, b) and cells lacking MRP5 expression (Figure 2 c, d), the binding of ML-I to both cell types is absent after periodate treatment (Figure 2 b, d). The co-localization of MRP5 and ML-I is shown in Figure 3.
Binding of peptides to ML-I/competition of ML-I binding. Histochemical binding with (a) 100 mM ML-I alone or in combination with (b) 2 mM, (c) 5 mM, or (d) 10 mM of the peptide HSSWWLALAKPTC, or with (e) 10 mM, (f) 20 mM, or (g) 50 mM D-galactose (control inhibitor of ML-I).
ML-I binding to suger chains of the MRP5 molecule. Histochemical binding assays of ML-I binding to MRP5-overexpressing cells (a, b) and lacking MRP5 cells (c, d), without periodate treatment (a, c) and with periodate treatment (b, d).
MRP5 and ML-I double staining. Histochemical binding assays showing fixed (a, b) and native (c, d) MRP5-overexpressing 293 HEK cells and cells lacking MRP5. ML-I binding to MRP5-overexpressing cells and lacking MRP5 cells.
CMFDA efflux in control (a) and MRP5-overexpressing (b) 293 HEK cells. One of three independent experiments. The mean fluorescence is denoted in arbitrary units (au).
Effect of sulfinpyrazone on CMFDA-exclusion by MRP5 overexpressing cells. (a) without blocker, (b) sulfinpyrazone present at 1 mM. The mean fluorescence is denoted in arbitrary units (au).
Effect of mistletoe lectins on CMFDA exclusion by MRP5 overexpressing cells. (a) Control, non-transfected 293 HEK cells, (b) MRP5 overexpressing 293 HEK cells alone or with (c) ML-I, (d) ML-II, (e) ML-III present at 10 μg/ml. The mean fluorescence is denoted in arbitrary units (au). One of three independent experiments is illustrated.
Fluorochrome accumulation/efflux by HEK293 cells overexpressing MRP5. In three independent experiments the mean fluorescence value in the non-transfected 293 HEK cells was 7057 arbitrary units (au) (range 5923-7649) compared with 2060 au (range 1866-2341) in the MRP5 transfected cells. As shown in Figure 4 CMFDA-efflux increased in the transfected cells. To confirm that the efflux of CMFDA observed in the MRP5-transfected cells was due to MRP activity, the MRP5 blocker sulfinpyrazone was used resulting in the retention of high fluorescence as shown in Figure 5. In two independent experiments, the mean fluorescence without sulfinpyrazone was 2913 au compared with 7695 au with sulfinpyrazone, indicating that the CMFDA exclusion was due to MRP5.
Modulation of fluorochrome efflux by MLs. All three MLs induced activation of MRP5, which was seen in the reduced fluorochrome labelling of the cells (Figure 6). The most effective activation was demonstrated by ML-II, whereas ML-I and −III only showed a slight increase of fluorochrome efflux. The mean fluorescence value in the MRP5-transfected 293 HEK cells without MLs was 2060 au (range 1866-2341), with ML-I was 2048 au (range 1609-2577), with ML-II was 1912 au (range 1102-2436) and with ML-III was 2043 au (range 1462-2455). There was no effect of the three MLs on CMFDA efflux in non-transduced HEK cells (data not shown).
Discussion
Various peptide sequences binding to MLs were identified. Upon sequence analysis, the sequence of the binding peptides showed homologies to various intracellular and membrane proteins. As it is well established that carbohydrate residues recognised by lectins are limited to the extracellular surface of the cell membrane of intact living cells (30), focus was directed to peptide sequences of membrane proteins which could serve as putative binding sites for therapeutically administered MLs. Because the sequence from MRP5 produced significant alignments in the BLAST search with the HSS sequence, this transporter channel was focused upon. MRP5 is an organic anion transporter of glutathione conjugates and MRP5 can be inhibited by typical organic anion transport inhibitors. As the fluorochrome CMFDA is a suitable substrate for MRP5, a direct measurement of the functional activity of the MRP5 dye efflux pump in living cells could be performed under the influence of the three MLs and their pharmacological action with regard to this pump could be observed. Although the MLs modulated the efflux pump slightly, the effect is probably biologically insignificant. These minimal effects ranging from 1 to 6% seem to be biologically meaningless if one considers that the MRP5 proteins are strongly overexpressed due to the genetic enginering. This was in contrast to observations with tomato lectin, which had a considerable influence on the function of the multidrug resistance protein 1 (MDR-1) exclusion pump [33].
While the HSS peptide represents a sequence of the MRP5 protein, it might therefore be assumed that ML-I binds to this peptide sequence within the MRP5 molecule. However, as MRP5 is also glycosylated, it would also be possible that ML-I binds to a carbohydrate residue on MRP5. To investigate whether ML-I binds to a glycan on MRP5, periodate treatment was performed. As periodate treatment abolished the ML-I binding sites completely in the MRP5 overexpressing cells, the binding of ML-I to the MRP5 overexpressing cells could most likely be attributed to the sugar chains of the MRP5 molecule and other galactose-binding glycoproteins as well. As no residual ML-I reactivity was observed, binding of ML-I to the HSS sequence in the MRP5 protein, even in MRP5 overexpressing cells, cannot be postulated. Thus, despite the presence of both peptide- and sugar-binding sites for ML-I, only the latter seem to be of importance for ML binding. Thus, no protein sequences of functional importance for the binding of MLs were identified. The diverse biological effects mediated by the MLs must therefore originate from lectin-sugar interactions.
Footnotes
-
↵* Both authors contributed equally to this article.
- Received August 6, 2009.
- Revision received October 29, 2009.
- Accepted November 4, 2009.
- Copyright© 2009 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.